If it ain't broken, don't fix it (but if it is, make sure you know): aortic valve interventions during left ventricular assist device implantation.

Abstract

This article refers to ‘Survival following a concomitant aortic valve procedure during left ventricular assist device surgery: an ISHLT Mechanically Assisted Circulatory Support (IMACS) Registry analysis’ by J.F. Veenis et al., published in this issue on pages 1878–1887. Incompetence of the aortic valve constitutes a haemodynamic challenge for patients implanted with a continuous-flow left ventricular assist device (LVAD). Aortic regurgitation (AR) results in a short-circuit circulation between the ascending aorta, the left ventricular (LV) cavity and the pump, resulting in inadequate LV unloading and reduced systemic cardiac output, despite a well-functioning device.1 AR after LVAD implantation is associated with reduced functional capacity and increased mortality.2, 3 Uncorrected moderate preoperative AR is likely to progress over time and consequently European and international guidelines clearly recommend correction of moderate or severe AR at the time of implant.4 Repair, if possible, or replacement of the incompetent aortic valve with a bioprosthesis are the interventions of choice. Insertion of a mechanical valve is distinctly discouraged as it is likely to be dysfunctional and subject to thrombosis because of the continuous flow. For the same reasons, a pre-existing mechanical valve, even if well-functioning, should be replaced with a bioprosthesis.4 In rare cases, over-sewing of the aortic valve is performed but is recommended only if potential for recovery of the left ventricle is ruled out and other options are impossible.5 Guidelines clearly recommend thorough examination of the aortic valve using transthoracic or transesophageal (TEE) echocardiography prior to LVAD implantation.4, 6 Simultaneous aortic interventions increase the complexity of the LVAD implant procedure, may limit the surgical access options and prolong cardiopulmonary bypass time. Consequently, it is of vital importance to understand how AR and the need for perioperative aortic valve intervention affect outcomes. This information is crucial to discuss risks with patients and – ultimately – to identify high-risk patients where AR ‘on top’ of other comorbidity might confer a risk so high that LVAD implantation should be avoided. Early data from a large cohort of patients undergoing HeartMate II implantation clearly identified the need for aortic valve intervention as a clinically very significant risk factor for 6-month mortality in contrast to other cardiac surgery interventions performed simultaneously with the LVAD implantation (for instance closure of a patent foramen ovale or coronary revascularization).7 However, expertise has grown, devices have become smaller and allow for shorter cardiopulmonary bypass time. Hence it is important to revisit this issue to provide the most accurate and current information to guide cardiologists and surgeons in assessing the risk of patients requiring aortic valve interventions at the time of LVAD implantation. In the current issue of the Journal, Veenis et al.8 used data from the International Society of Heart and Lung Transplantation Mechanically Assisted Circulatory Support (IMACS) registry to analyse how AR and aortic valve interventions affected outcome in 15 267 recipients of a primary continuous-flow LVAD. Of these, 785 (5%) underwent a concomitant aortic valve procedure at the time of LVAD implantation. Patients undergoing aortic valve interventions were older and more often implanted as destination therapy, but not very different with respect to other risk factors from patients not undergoing aortic valve replacement or repair. This suggests that patients known to require concomitant aortic valve surgery were generally selected carefully. Overall, patients undergoing aortic valve interventions had lower survival compared with those with aortic valve interventions even after adjustment for other risk factors. The excess risk associated with aortic valve intervention was driven by a difference in the early postoperative period, and the magnitude of the unadjusted difference was modest (90-day survival 90.4% vs. 85.1% and 87.4% in patients undergoing aortic valve replacement or repair, respectively). Interestingly, the increased mortality after combined LVAD and aortic valve intervention was confined to the group of patients without documented moderate-severe AR at ‘baseline’, which must be interpreted as AR being recognized prior to LVAD surgery. Again, this group was surprisingly large as it constituted 433 patients (55% of the patients with aortic valve interventions). Hence more than half the patients undergoing either aortic valve replacement or repair had none or mild AR recorded at baseline. This information appears counterintuitive. Why would patients undergo aortic valve surgery if moderate-severe AR had not been documented? Unfortunately, the indication for the procedure was not known. First, a minority of patients might have had aortic stenosis (with no or mild AR) complicated by severe LV failure requiring LVAD implantation. No data to estimate this population are available in the study but it is likely to be small. Also, in some cases aortic valve replacement of a competent mechanical prosthesis with a bioprosthesis has likely been performed, as recommended in the guidelines. However, most likely, in a significant proportion of the patients undergoing aortic valve interventions without documented moderate or severe AR at baseline, significant valvular incompetence was not realized until during the LVAD operation. In patients with severe LV failure and low systemic blood pressure – the typical LVAD candidate – AR may easily be underestimated, but becomes apparent after LVAD implantation or during cardiopulmonary bypass as the systemic blood pressure is increased by the mechanical circulation throughout the entire cardiac cycle due to the continuous flow. In some patients significant AR may simply have been overlooked if adequate echocardiography – TEE in the case of imperfect visualization of the aortic valve on transthoracic views – has not been performed. In patients with a diagnosis of moderate or severe AR made in the operating room, addition of an unplanned cardiac procedure after the LVAD implant may substantially increase the risk of adverse outcome. In some patients, knowledge about the need for simultaneous aortic valve intervention might have caused re-evaluation of a patient's candidacy for LVAD implantation or consideration of other treatment strategies, for instance implantation of a total artificial heart. With the data granularity offered by IMACS, unfortunately we cannot determine the underlying reasons for the large number of interventions in patients without known moderate-severe AR in the study, but given the apparent interaction of degree of AR on the effect of aortic interventions on short-term mortality, we speculate that significant AR discovered on TEE only in the operating room surprising the LVAD surgical team is mainly responsible. The IMACS data suggest that more studies are necessary to determine the optimal approach in this situation. The data of the present analysis confirm the need for a thorough echocardiographic assessment before surgery to identify patients with moderate to severe AR. In contrast, patients with no or mild AR, confirmed at the time of surgery, should not undergo aortic valve replacement, which is also supported by the data from the IMACS study. However, these patients should be followed life-long with echocardiography at regular intervals to assess the competency of the valve as AR may develop or progress, especially those with mild AR, and may require intervention later.3 Another surprising finding of the study was the significant number of patients in whom a mechanical valve was chosen for aortic valve replacement (16% of aortic valve prostheses implanted) despite the guideline recommendations against this approach.4 Not surprisingly, these patients had the worst survival, and this should serve as a reminder to avoid this strategy. A suggestion on how to manage AR in patients scheduled for LVAD implantation based on the current knowledge is presented in Figure 1. New solutions to mitigate the problem of AR in LVAD recipients are evolving. Percutaneous transcatheter aortic valve interventions, such as transcatheter aortic valve replacements (TAVR) or catheter-based device closure of the aortic valve ostium (e.g. with an Amplatzer™ septal occluder), are being increasingly used in patients who develop AR after LVAD implantation.9 However, the experience with these strategies is still limited and complications such as valve migration or paravalvular leakage do occur. While hybrid TAVR and LVAD implantations have been described, there is little to support general use of this approach at the current time.10 In contrast, TAVR appears to be an attractive solution to patients developing AR later in the course during ongoing support with an LVAD, especially if they are high risk with conventional surgery, which is often the case.11 Aortic regurgitation remains a challenge in patients undergoing LVAD implantation. Clearly, more data are needed to understand when and how best to intervene in the incompetent valve. Randomized clinical trials would be extremely helpful in guiding the field, but to the knowledge of the authors no such studies are underway, leaving us with observational evidence. Despite its limitations, the study by Veenis et al. conveys two important points in this regard. First, moderate to severe AR diagnosed prior to LVAD implantation and managed appropriately with either repair or replacement with a bioprosthesis does not necessarily increase mortality risk significantly in carefully selected LVAD recipients. Second, and in contrast, it is possible that patients with AR requiring surgery which was not diagnosed prior to LVAD implantation have a poor survival as do patients where AR is corrected by a mechanical prosthesis. Proper preoperative investigation of the aortic valve remains crucial to plan the management of patients with advanced heart failure being considered for LVAD implantation. Conflict of interest: F.G. is Advisor to Abbott and Carmat. The other authors have nothing to disclose.